English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Recent computational approaches for combating online hate speech involve theautomatic generation of counter narratives by adapting PretrainedTransformer-based Language Models (PLMs) with human-curated data. This process,however, can produce in-domain overfitting, resulting in models generatingacceptable narratives only for hatred similar to training data, with littleportability to other targets or to real-world toxic language. This paperintroduces novel attention regularization methodologies to improve thegeneralization capabilities of PLMs for counter narratives generation.Overfitting to training-specific terms is then discouraged, resulting in morediverse and richer narratives. We experiment with two attention-basedregularization techniques on a benchmark English dataset. Regularized modelsproduce better counter narratives than state-of-the-art approaches in mostcases, both in terms of automatic metrics and human evaluation, especially whenhateful targets are not present in the training data. This work paves the wayfor better and more flexible counter-speech generation models, a task for whichdatasets are highly challenging to produce.

Weigh Your Own Words: Improving Hate Speech Counter Narrative Generation via Attention Regularization